An intelligent operation and maintenance system for wind power equipment based on domain-specific large models and multi-agent collaboration

Abstract

The application belongs to the technical field of intelligent operation and maintenance of wind power equipment, and relates to a wind power equipment intelligent operation and maintenance system based on a domain large model and multi-agent cooperation. The system comprises a multi-modal data acquisition module, a wind power pre-training representation / feature extraction module, an anomaly identification module, a wind turbine state knowledge graph and experience library, a tool chain calling decision module, an interpretable reasoning and verification module, a diagnosis report / work order generation module, and an online learning and feedback optimization module coupled with each module. The application constructs an intelligent operation and maintenance system of a "multi-agent cooperation + domain large model" architecture, and realizes unified scheduling and automatic execution of operation and maintenance tasks in combination with tool chain support, thereby significantly improving the autonomous decision-making capability and operation and maintenance efficiency of the system.

Images